Security issues session. Chris Anderson. SynBERC.

Environmental release and design precautions.

I am a professor at UC Berkeley. There's also a UCSF funded bioengineering center. And a bunch of other names that you've probably heard of. They asked one of us to come today and talk abuot our prespectives abotu issues and safety. I will probably talk about things that will freak you out. They are a little different. So what about biosafety in synthetic biology?

So far the discussion has been about synthesizing ricin or smallpox. You put in the sequence, you get it, you order it, and then you kill your buddies. I don't find that interesting. My field is about design. I haven't worried about people designing dangerous stuff is that it's really hard. Nothing has really changed since how things were done 30 years. We've gotten better at it, and unless you're super-educated, you're not going to make anything super dangerous.

But things are changing. And some things have come together in the past year, there are certain technologies fitting with old pieces, and there's a clear roadmap about what the next 5 or 10 years. And there are some biosafety threats emerging. Synthetic biology is going to be more software driven. You use robots more than your hands for constructing DNA. You use robots to determine if your designs are doings omething useful or not.

The motivations are due to the revolutions in sequencing, and what's coming in terms of synthesis, it's effecting all aspects of biology. The advances in sequencing is dramatic, and it's many of orders of magnitude than what we have imagined. So synthesis hasn't changed that much, it hasn't effected the bottom line. The microarray gene synthesis works, they are being packaged for commercial production and they will drop the price of genes down to about a factor of 10 within 5 years. You are going to construct sequences that you have mined from the metagenome from all the genome sequences, you're going to characterize the function of the gene, and you will somehow make that public for design.

BioFAB was the first. But now there's others like JGI from the DOE. They are shifting towards this fab mode of characteristics... all the genome sequences.. you're going to do bioinformatics stuff to figure out the 1000 thing within that that are worth characterizing. All sorts of people including myself are developing these multiplexing these that will allow 1000 genes per day, to figure out what is their molecular function, how they are likely to relate to their chases and other things. And hopefully biosafety characterization becomes a part of that. Hopefully that will be captured by software and distributed through registries.

Everyone is creating these software tools and wetlab tech that dovetails. You're doing a search with BLAST to get your list of genes, then we characterize enzymes, and some of the substrates hit on each enzyme, and each enzyme is its own story about what it does. But what has thus far happened in this field, this characterization part, is sometimes there's a datasheet, sometimes there's another step, and there's some CS researchers that have done encapsulation where you do, this is speciic observations about what these enzymes do.. but we care more about a general description about a pattern over which you can compute and design synthesis. There are some substrates.. we can define these patterns, not just the interactinos of small molecules.. so what about all this?

So that enables you to make tools that look like this. Where you say you want to make this chemical, and it just pops out a DNA sequence that is likely to make that product, or that it's not physically possible, or how to do it. This is the part that should freak you out. So there's a problem with this, you don't know- this might be open source. We don't know what people are going to type into this thing. Maybe they type in tetrodotoxin. That's probably not a good idea to tell them how to produce that. But then things can get weird.. what about TNT? Nobody wants to make tetrodotoxin, but maybe TNT. It's biologically active at small concentrations-it's a long pathway, it's hard, I'm not worried about it.

But I am more concerned about drugs, like methamphetamine. I think meth might be easy to synthesize at biologically active quantities with existing technologies. Why is this a problem? I'm not into drugs myself, but there are dangers. They are probably going to do it with a $40 yogurt maker, you do the engineering in a yeast... then there's a problem. This is someone who .. thoroughly deskilled at this point. It takes no effort to make this methcoli. Maybe she thinks it's about French culture, and then it grows in her intestines, and she gets either really high or really dead. This is a rusty nails situation. There is a risk that is not intuitively related to what you were doing.

This is what scares me. Somehow you have to mitigate that risk. There's also the drug thing with geopolitical consequences of being able to make drugs easily. Right now, why do we have all this international trade of drugs? You can't make any of the drugs we have right now in a clandestine way. You have to have giant tanks, or fields of poppeye. I have no idea what the effects of this are, but I think they are probably really big.

So what about mitigating the risks associated with this? What about if we just block all possible ways to do this? We could block the production of any of these dangerous things? We could have someone decide. But it would also block legit things, and people will work around it. Another thing we could do- we plan to introduce blocks of some sort, but we don't have a plan for how those blocks will work. There are people that have funded a production of biosafety strains. This is more to deal with the situation where the person ate the yogurt, and it bloomed and she died. You could make a strain of everything that is in some way broken, in the sense that it needs some small molecule in order to survive, and if you leave it out, it dies. There are so many efforts of these, I say it will exist, about 12 months at most, and we need to promote their use in the DIY community.

The third thing is the Clotho platform. Clotho is very active right now. They are building v3. Not only are you going to this portal for getting material to do design work and such, it's also going to have a site where things can link out to for educational resources, so there are other tools. You recognize it's dangerous, and it pops you to an education page, you have to take this quiz at the end, and make suer you know what it is, and if it's not really bad we'll give you a certificate and knowledge of how to make it.

There's already one guy-......

Jeff Sampson Phillip Elms Sugiarto Go Tim Hsiau Nima Emami Trent Northern Douglas Densmore